Question

I have the following table

drop table names;
create table names(
    name    varchar(200),
    surname varchar(200),
    primary key(name, surname)
);

insert into names values ('john', 'doe');
insert into names values ('john', 'richards');

I’d like to list the surnames where the name is the same.
i.e. I want this result

----------+----------
 doe      | richards
 doe      | doe
 richards | richards

(this is only an example, the real problem is quite different)

The followig query returns duplicates

select n1.surname, n2.surname
from names n1 join names n2
on n1.name = n2.name;

 surname  | surname  
----------+----------
 doe      | richards
 doe      | doe
 richards | richards
 richards | doe

So, I used the least and greatest postgresql functions, and got my result. But: least and greatest are not standard SQL, and I was wondering whether a more efficient solution exists, as using a join condition I cannot imagine now.

select distinct least(n1.surname, n2.surname), greatest(n1.surname,
n2.surname)
from names n1 join names n2
on n1.name = n2.name;

  least   | greatest 
----------+----------
 doe      | richards
 doe      | doe
 richards | richards

Answer 1

If you only want a tie-breaker, use < of > or >= or <=

SELECT n1.surname, n2.surname
FROM names n1 
JOIN names n2 ON n1.name = n2.name
WHERE n1.surname >= n2.surname
   ;